From:	efiring AT iniki DOT soest DOT hawaii DOT edu (Eric Firing)
Subject:	Math coprocessor support in 17.1?
23 Dec 1996 19:41:37 -0800 :
Sender:	daemon AT cygnus DOT com
Approved:	cygnus DOT gnu-win32 AT cygnus DOT com
Distribution:	cygnus
Message-ID:	<199612240251.QAA11630.cygnus.gnu-win32@ahi.soest.hawaii.edu>
Original-To:	gnu-win32 AT cygnus DOT com
Original-Sender:	owner-gnu-win32 AT cygnus DOT com

I just installed 17.1, and compiled and ran a short
benchmark.  Execution times were compared to those from
DJGPP and Turbo C versions.  Cygwin32 did fine so long as the
benchmark did not include a floating point op, but putting in
calls to sin() slowed it down enormously in comparison to both
other compilers.  It appears that the floating point ops are
being done in software, not in the FP coprocessor.  (My machine
is a Toshiba 420, with Pentium 100.  OS is Win95.) 

Is the FP coprocessor supported at all?  If so, how can I ensure
it is used in place of the softare emulation that seems to be
occurring by default?

Thanks.

Eric Firing
efiring AT soest DOT hawaii DOT edu
-
For help on using this list, send a message to
"gnu-win32-request AT cygnus DOT com" with one line of text: "help".

- Raw text -